The description of the mechanisms underlying the transformation of synchronic language variation into diachronic change (Gardiner & Nagy 2017; Sankoff & Blondeau 2007; Labov 1994; Weinreich et al. 1968) and the influence on language evolution of competing synchronic pressures in language production and perception (e.g. Wedel 2012; Bybee 2007; Martinet 1955; Zipf 1949) are topics of longstanding interest in the study of language change. Constraints of this sort can be perceived through their realization on particular concrete forms or structures. In this dissertation, I present three discrete case studies organized into article-like chapters, each of which examines the interaction of form and function in synchrony and diachrony, fitting the drivers behind each scenario of language change to a particular explanatory framework.
Chapter 2 refines our understanding of the role of functional load (Surendran & Niyogi 2006, 2003; Martinet 1955; Hockett 1955) in the resistance to merger of certain phonemic contrasts (Wedel et al. 2013). Building on Wedel et al. (2013) by incorporating word co-occurrence vectors (Mikolov et al. 2013), we provide evidence that phonemic contrasts in minimal pairs of high contextual similarity are less likely to undergo phonemic merger, corroborating the finding of Wedel et al. (2013) for words that share discrete parts-of-speech without requiring the a priori determination and annotation of word classes.
Chapter 3 examines asymmetrical patterns of sound change within words through the lens of a theory of speech perception, the cohort model of lexical access (Marslen-Wilson & Welsh 1978), which I argue subjects different subwords to differing cognitive processes and, therefore, potentially-heterogenous selection pressures. By analyzing cognate pairs across six Indo-European languages and dividing them into intralexical subregions predicted by cohort theory, I corroborate earlier work by showing that, in pairs of cognates, phonemes that occur relatively further inside words are more likely to undergo change (Wedel et al. 2019; Brendel 2018). Additionally, I suggest that, when recognition points are incorporated into the analysis, material after a recognition point has a pattern of predicted change that differs from material subjected to cohort selection in addition to a general end-of-word tendency to change, suggesting that the unequal deployment of the cognitive machinery of perception shapes the long-term trajectory of sound change at a sublexical level.
In the historical case study presented in Chapter 4, formal inadequacies and system-level 'preferences' interact during a centuries-long period of variation and instability (as in Nichols 2003) in the relativization system of Early Modern Icelandic. Through a study of diachronic written corpora (Wallenberg et al. 2011), I examine the interactional, system-level, and sociocultural factors which conspired to maintain a preference for a certain strategy of relativization despite extensive contact pressure, ongoing instability in the relevant grammatical subsystem, and formal change (lexical replacement). This story ends in restored equilibrium after instability, and I suggest a sort of system elasticity or inertia, perhaps most readily observable across time, that can exert pressures on synchronic realizations.
Taken together, these chapters contribute to our understanding of how languages bear out competing pressures—variously interactional, cognitive, and, in the final chapter, sociocultural—while maintaining systemic coherence over time.
